Enzyme–Substrate Binding Kinetics Indicate That Photolyase

Article pubs.acs.org/biochemistry

Enzyme−Substrate Binding Kinetics Indicate That Photolyase Recognizes an Extrahelical Cyclobutane Thymidine Dimer Johannes P. M. Schelvis,*,† Xuling Zhu,‡,§ and Yvonne M. Gindt† †

Department of Chemistry and Biochemistry, Montclair State University, 1 Normal Avenue, Montclair, New Jersey 07043, United States ‡ Department of Chemistry, New York University, 100 Washington Square East, New York, New York 10003, United States ABSTRACT: Escherichia coli DNA photolyase is a DNA-repair enzyme that repairs cyclobutane pyrimidine dimers (CPDs) that are formed on DNA upon exposure of cells to ultraviolet light. The light-driven electron-transfer mechanism by which photolyase catalyzes the CPD monomerization after the enzyme−substrate complex has formed has been studied extensively. However, much less is understood about how photolyase recognizes CPDs on DNA. It has been clearly established that photolyase, like many other DNA-repair proteins, requires flipping of the CPD site into an extrahelical position. Photolyase is unique in that it requires the two dimerized pyrimidine bases to flip rather than just a single damaged base. In this paper, we perform direct measurements of photolyase binding to CPD-containing undecamer DNA that has been labeled with a fluorophore. We find that the association constant of ∼2 × 106 M−1 is independent of the location of the CPD on the undecamer DNA. The binding kinetics of photolyase are best described by two rate constants. The slower rate constant is ∼104 M−1 s−1 and is most likely due to steric interference of the fluorophore during the binding process. The faster rate constant is on the order of 2.5 × 105 M−1 s−1 and reflects the binding of photolyase to the CPD on the DNA. This result indicates that photolyase finds and binds to a CPD lesion 100−4000 times slower than other DNA-repair proteins. In light of the existing literature, we propose a mechanism in which photolyase recognizes a CPD that is flipped into an extrahelical position via a three-dimensional search.

P

ments.15−18 A later X-ray crystal structure of photolyase complexed to CPD-containing DNA confirmed the extrahelical binding geometry, although the CPD appeared to have been repaired during the data collection process.19 Photolyase clearly requires an extrahelical CPD binding geometry, but NMR solution and X-ray crystal structures of CPD-containing DNA show an intrahelical CPD;20,21 therefore, the CPD has to flip out of the DNA helix for the formation of the enzyme− substrate complex. Although other DNA-repair proteins also require base flipping of the damaged site,22,23 photolyase is unique because it requires two dimerized thymidines to flip out of the double helix rather than a single damaged base. The role of the enzyme in base flipping is not completely clear. A computational study of csCPD-containing DNA and a recent isothermal calorimetry study of photolyase−substrate binding provided support for binding of photolyase to an extrahelical CPD that spontaneously flips out of the double helix.24,25 However, a recent NMR study of base flipping in csCPDcontaining DNA argued against the spontaneous movement of the csCPD into an extrahelical position as a viable mechanism.26 Over the past few decades, a general picture of how proteins search for and find a specific sequence or a specific site target

hotolyases are structure-specific DNA-repair enzymes that are found in all kingdoms of life except for placental mammals, and they repair DNA lesions that have been induced by ultraviolet (UV) light.1−3 The most common photolyase repairs the major DNA UV photoproduct, the cis,syn cyclobutane pyrimidine dimer (csCPD) that forms between adjacent pyrimidine bases.1,4,5 The repair mechanism can be broken down into two separate but equally important steps. First, the enzyme searches for and binds to the damaged site on DNA. Second, once the enzyme−substrate complex has formed, the enzyme has to absorb a blue photon to initiate the catalytic cycle of csCPD monomerization. While the mechanism by which DNA photolyase repairs CPD lesions after formation of the enzyme−substrate complex is relatively well understood,1,6−8 the means by which the enzyme recognizes and binds to CPD lesions on DNA is still not clear. The first X-ray crystal structure of DNA photolyase indicated that the active site flavin adenine dinucleotide (FAD) was located at the bottom of a cleft that was presumed to be the CPD binding site.9 This structural evidence suggested that the CPD would have to assume an extrahelical position to bind to DNA photolyase. Subsequently, such a binding geometry was predicted by computer modeling and computational studies.10−12 The first experimental evidence of extrahelical CPD− photolyase binding geometry came from the Stanley group,13,14 and their interpretation was supported by subsequent nuclear magnetic resonance (NMR) and optical spectroscopy experi© 2015 American Chemical Society

Received: August 19, 2015 Revised: September 14, 2015 Published: September 22, 2015 6176

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry Table 1. Properties of the Oligonucleotides Used in This Study molar extinction coefficient at 260 nm (M−1 cm−1) oligonucleotide Oligo1 Oligo2 Oligo3 Complementary-1 Complementary-2 Complementary-3 a

sequence (5′ → 3′) GCA CGA GTT CTC CTC GTC

a

AGT TGG AG TTG GAG AG GGA GAG AC CAA CTT GC TCC AAC TG TCT CCA AC

b

c

without CPD

with CPD

113300 114400 114600 93800 94900 96900

97900 99000 99200

molar extinction coefficient at 521 nm (M−1 cm−1) with TET at the 5′-terminusb

86000 86000 86000

Position of T-T to form cis,syn CPD shown in bold. bData from TriLink Technologies. cCalculated by using the nearest-neighbor model.42

on DNA has emerged.27,28 First, the protein binds nonspecifically to the DNA with a diffusion-controlled rate constant. Then, the protein uses “facilitated diffusion” to locate its target sequence or site on the DNA. The facilitated diffusion can occur through sliding, hopping, or jumping to search for the target in a unidimensional way. During sliding, the protein fully remains in contact with the DNA, but it is only efficient over short distances, 10−50 bp.27,28 During hopping (100 bp), the protein dissociates from the DNA molecule but remains in its vicinity because of electrostatic interactions and lands on a different part of the same DNA molecule. All three processes will compete with the complete dissociation of the protein from the DNA. From the literature, it appears that the proteins that follow this search mechanism bind to their target on the DNA with a binding rate constant ranging from 1.7 × 107 to 9.0 × 108 M−1 s−1,29−34 including proteins that require base flipping as part of the search process. Indirect determination of the binding rate constant of photolyase in Escherichia coli cells in vivo and with plasmid DNA in vitro gave rate constants of 1.1 × 106 and 4.2 × 106 M−1 s−1, respectively, at room temperature.35,36 These rate constants are 2−3 orders of magnitude slower than those reported for other site-specific DNA-repair enzymes29−34 and hint that the DNA damage recognition and binding process is quite different and potentially more complex for photolyase. Direct measurements of the kinetics of the formation of the enzyme−substrate complex are required to address this discrepancy. In this work, we investigate the binding of photolyase to csCPD-containing dsDNA as a function of CPD position on the DNA by using steady-state and stopped-flow fluorescence spectroscopy. The DNA is tagged with a fluorophore, tetrachlorofluorescein (TET), and E. coli photolyase is used in its purified form with the flavin cofactor present as the neutral radical semiquinone (FADH•).37 Under these conditions, the binding of photolyase to csCPD-DNA can be directly monitored by the quenching of the TET excited state via Förster resonance energy transfer (RET) to the FADH• in photolyase. Förster RET between two chromophores has three main requirements: the donor−acceptor distance should be 85% of the decay of the fluorescence intensity. In contrast, the kinetic traces for csCPD-DNA2 are best described by a monoexponential decay function with the rate of binding of photolyase to csCPD-DNA2 found to be similar to that of the slow phase that is observed for csCPD-DNA1 and csCPD-DNA3. We did not find any evidence of faster kinetic signals. First, the percent change in fluorescence intensity was virtually unchanged: 25 ± 1% for DNA1 and 17 ± 1% for DNA3 at the various enzyme concentrations. Second, kinetic traces taken with a 1 s run time and 0.5 ms integration intervals and with a 10 s run time and 5 ms integration intervals are identical within the noise of the experiment. The insets in Figure 3 show the apparent rate constants plotted as a function of enzyme concentration (■) together with a linear fit (). The rate constants for binding of photolyase to the three different csCPD-containing DNAs are determined by the slope of the linear fit with assumption of first-order enzyme kinetics and shown in Table 3. The rate

Figure 1. Absorbance-detected thermal denaturation curves of (A) DNA1, (B) DNA2, and (C) DNA3 in the absence () and presence (···) of the TET fluorophore and in the absence or presence of the csCPD as indicated in the graph. The absorbance change of the denaturation curves has been normalized to facilitate comparison.

likely due to fluorescence quenching by the GC base pair adjacent to TET.46 Addition of a 3-fold excess of photolyase to csCPD-DNA1 causes an additional 13% decrease in TET fluorescence (Figure 2A, dotted line). This decrease in TET fluorescence is due to RET from TET to the FADH• in photolyase, as described above. The emission spectra for TET-labeled complementary strands of Oligo2 and Oligo3 are shown in panels B and C of Figure 2, respectively (solid lines). Upon addition of csCPDOligo2 to form csCPD-DNA2, the TET fluorescence decreases by 53% (Figure 2B, dashed line), and addition of a 3-fold molar excess of photolyase to csCPD-DNA2 decreases the fluorescence by an additional 4% (Figure 2B, dotted line). In 6179

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry

DNAs was monitored by determining the melting temperature (Tm). Because the GC content was the same for all three DNA constructs, their melting temperatures were very similar, and the presence of the TET fluorophore at the 5′-terminus of the complementary strand did not change the Tm of the DNA constructs in a manner independent of the csCPD. Thus, we conclude that the presence of TET does not alter the thermal stability of the dsDNA constructs with or without the csCPD present in any significant way. The Tm of DNA1 is significantly higher (54 °C) than that reported in the literature (45 °C).44 The higher Tm determined in our studies is most likely due to a higher oligonucleotide concentration (3.0 μM compared to 0.70 μM), though a difference in the ionic strength of the oligonucleotide solutions could also have contributed to the observed difference in Tm. The effect of the csCPD on the thermal stability of the dsDNA constructs is significant and not unexpected. We observe differences in the Tm of 5−10 °C when undamaged and damaged DNA are compared (Table 2). This is in good agreement with the range of 6−9 °C that has been reported for other dsDNA constructs with 8−12 bp.20,43−45 For unlabeled DNA, the Tm of the undecamers is 53−54 °C, and the presence of the csCPD decreases it by 5 °C for DNA1 and DNA3 and by 9 °C for DNA2. In DNA1 and DNA3, the csCPD is flanked by GC base pairs on either side, whereas it is flanked by one GC (3′ of csCPD) and one AT (5′ of csCPD) base pair in DNA2. Work by the Stanley group on melting temperatures of csCPDdsDNA with fluorescent bases has shown that the base pairs on the 5′-side of the CPD, including the 5′-T of the csCPD, are slightly less stable than those on the 3′-side of the CPD.44,45,47 These results are corroborated by a recent NMR study that shows that the 5′-TA base pair of the CPD is more labile than its 3′-TA base pair.26 Therefore, it seems that the flanking 5′TA base pair of the csCPD in DNA2 destabilizes the csCPDdsDNA slightly more than a flanking GC base pair. This change results in a larger decrease in the Tm of DNA2 in the presence of the csCPD. Because the Tm is the same for DNA1 with the csCPD in the center and DNA3 with the csCPD 1 bp from the 5′-terminus, it appears that end fraying near the csCPD in DNA3 does not result in an additional decrease in the Tm. Influence of the Location of the CPD on the Binding of Photolyase. The location of the csCPD on the undecamer dsDNA seems to have only a minor effect on the association constant, KA, of DNA photolyase; KA is on the order of 2 × 106 M−1. Although the csCPD in DNA-3 is only 1 bp from the 5′terminus, this location does not appear to impede the binding process. All three DNA constructs have the minimal binding unit, i.e., one phosphate 5′ and three phosphates 3′ of the csCPD.48 The KA value is significantly lower than what has been published for plasmid DNA (KA = 4.7−14 × 107 M−1), a 43mer dsDNA (KA = 108 M−1), and poly(dT) (KA = 4.8 × 108 M−1), each containing one or more csCPDs,36,49,50 while measurements on UV-p(dT)2 and UV-p(dT)4 gave association constants of 1.8 × 104 and 4.0 × 106 M−1, respectively.51,52 Our measurements are in excellent agreement with a recent isothermal calorimetry (ITC) experiment, in which the equilibrium constant of binding of photolyase to csCPDcontaining poly(dA·dT)10 was determined to be 1.1 × 106 M−1 at 17 °C compared to our value of 2 × 106 M−1 at 15 °C.24 Our fluorescence method is validated because we obtain virtually the same binding constant with our TET-labeled DNA as that obtained from unlabeled DNA via ITC. The discrepancy between our value and those in the earlier literature has been

Figure 2. Fluorescence-detected binding of photolyase to csCPDcontaining DNA. Each graph shows the emission spectrum of the complementary strand with TET attached (), of the complementary strand with TET and a 1.4-fold molar excess of csCPD-containing oligonucleotide (---), and of the complementary strand with TET, a 1.4-fold molar excess of csCPD-containing oligonucleotide, and a 6fold molar excess of photolyase (···). Emission spectra are shown for (A) csCPD-DNA1, (B) csCPD-DNA2, and (C) csCPD-DNA3. The inset of each graph shows the photolyase titration curve with the inverse change in emission intensity on the y-axis.

constants for binding of photolyase to csCPD-DNA1 and csCPD-DNA3 are very similar, ∼2.5 × 105 M−1 s−1, and the rate constant for binding to csCPD-DNA2 is ∼10 times slower, ∼2.3 × 104 M−1 s−1.

■

DISCUSSION Influence of the Fluorophore and CPD on DNA Thermal Stability. The thermal stability of the csCPD6180

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry

Table 3. Equilibrium Association Constants and Association and Dissociation Rate Constants of the dsDNA Constructs Used in This Study

a

sequence

KA (×106 M−1)

kon,fast (×105 M−1 s−1)

kon,slow (×104 M−1 s−1)

koff (×10−2 s−1)a

5′-GCAAGTTGGAG-3′ 3′-CGTTCAACCTC-5′ 5′-CGATTGGAGAG-3′ 3′-GCTAACCTCTC-5′ 5′-GTTGGAGAGAC-3′ 3′-CAACCTCTCTG-5′

3.2 ± 0.4

2.5 ± 0.4

0.5−1.4

7.8 ± 2.2

1.9 ± 0.3

−

2.3 ± 0.3

1.2 ± 0.3

2.2 ± 0.3

2.4 ± 0.3

0.3−2.1

11 ± 3

Calculated by using the dominant rate constant; koff = kon/KA.

discussed in detail by one of us,24 and we will not elaborate further here. Given the kinetic traces shown in Figure 3, the position of the csCPD seems to have a significant effect on the rate of binding of DNA photolyase. In DNA1 and DNA3, the binding kinetics show both a fast phase and a slow phase, with the fast phase more dominant in DNA3 (≥85%) than in DNA1 (≈64%). The rate constant associated with the fast phase is ∼2.5 × 105 M−1 s−1 in both DNA1 and DNA3, and the slow phase is from 0.03 to 0.21 × 105 M−1 s−1. The binding kinetics of DNA2 has only a slow phase of 0.23 × 105 M−1 s−1, slightly faster than that observed for the slow phase in DNA1 and DNA3. The kinetics of the binding is monitored by the fluorescence of the TET fluorophore. The steady-state data (Figure 2) clearly show quenching of this emission upon protein binding to the DNA molecule, because of RET. The observation of two phases in our data indicates that we may have a complicating factor to consider; we may be detecting a steric effect due to the bulky fluorophore. Therefore, the observation of two rate constants appears to reflect two separate processes: a fast-phase unhindered binding step to photolyase and a slow step due to interference from the TET fluorophore. Our hypothesis is that the csCPD and the fluorophore are aligned in DNA2 in such a way that the fluorophore creates maximal steric interference with photolyase binding and only the slow phase is observed. When the csCPD is moved 2 bp either to the center of the DNA (DNA1) or to the terminus of the DNA (DNA3), the position of the csCPD rotates by approximately 72° and −72° with respect to the fluorophore, respectively, due to the helical nature of dsDNA. This rotation reduces the probability of steric interference of the fluorophore upon binding of photolyase to the csCPD. The observation that the slow phase is much smaller when photolyase binds to DNA3 than to DNA1 supports this hypothesis. In DNA3, the csCPD is farther removed (4 bp, ∼13 Å) from the TET fluorophore than in DNA1. The increased distance between the csCPD and TET in DNA3 most likely results in less steric interference of TET during binding, which then translates into a much smaller contribution of the slow phase to the binding kinetics of photolyase and DNA3. Therefore, fluorescent probes can be used to monitor the binding of photolyase to csCPD-DNA, but one has to pay careful attention to possible steric interference of the fluorophore in the binding process. In addition, the observation that KA is independent of csCPD location also suggests that the photolyase−substrate dissociation rate is slower with DNA2 than with DNA1 and DNA3. The dissociation rates have been calculated from koff = kon/KA and are listed in Table 3. Apparently, the fluorophore not only slows the formation of the photolyase−substrate complex in the case of DNA2 but also slows the dissociation of the complex.

Only two rate constants, both obtained by indirect measurements, for the binding of photolyase to csCPD-DNA have been reported in the literature. In 1970, Harm reported rate constants of 3.1 × 105 M−1 s−1 (5 °C), 1.1 × 106 M−1 s−1 (23 °C), and 2.1 × 106 M−1 s−1 (37 °C) for in vivo experiments.35 Sancar and co-workers determined binding rate constants of 1.4 × 106 and 4.2 × 106 M−1 s−1 at 25 °C for photolyase binding to CPD-containing plasmid DNA.36 From Harm’s work, one can calculate an activation energy of 46 kJ/ mol for formation of the enzyme−substrate complex. Using the Arrhenius equation with this activation energy, we calculate the rate constant at 15 °C to be 6.5 × 105 M−1 s−1 for Harm’s data and 7.3 × 105 to 2.2 × 106 M−1 s−1 for Sancar’s data. Our work shows that photolyase binds to the doublestranded undecamer csCPD-DNA constructs with rate constants that are 2.5−9-fold smaller at 15 °C than those reported for in vivo cellular and in vitro plasmid DNA studies. It is possible that this reflects a real difference in the binding kinetics of photolyase due to the length of the dsDNA. However, in the earlier work, the binding rate constants were determined indirectly, so some assumptions and estimates to extract the binding rate constants from the experimental data were required; these assumptions introduce some uncertainties into the earlier results. However, it should be noted that neither we nor Sancar finds evidence for binding rate constants on the order of 108 M−1 s−1 as has been determined for other DNArepair systems. Now that we have a direct way to measure the binding rates of photolyase to its substrate, we can design experiments to investigate inconsistencies with earlier work. Implications for the Mechanism of CPD Recognition by DNA Photolyase. We determined a rate constant of 2.5 × 105 M−1 s−1 for E. coli photolyase binding to the site-specific CPD lesion on dsDNA. This rate constant is significantly smaller than what has been reported in the literature for both sequence-specific and other site-specific DNA binding proteins. For example, the sequence-specific Lac repressor has a binding rate constant of approximately 108−109 M−1 s−1 at physiological ionic strengths.29,30 The site-specific DNA binding enzymes uracil DNA glycosylase (UDG) and DNA methyltransferase M.EcoRI bind to their target with rate constants of 2−4 × 108 and 1.7−5.0 × 107 M−1 s−1, respectively, while Xenopus XPA binds to a fluorescein-labeled site with a rate constant of 9.0 × 108 M−1 s−1.31−34 In general, these large rate constants are attributed to diffusion-controlled nonspecific binding of the protein to DNA followed by one-dimensional facilitated diffusion through sliding, hopping, and/or jumping of the protein in the proximity of and along the DNA to locate and bind to the target sequence or site.27,28 On the basis of their photolyase binding studies, Sancar and co-workers estimated a ratio of specific to nonspecific binding of photolyase of 2 × 105, and they calculated that electrostatic 6181

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry

approach to damage recognition than other DNA-repair systems studied. Photolyase and the undecamer dsDNA will have similar three-dimensional diffusion coefficients and, therefore, a diffusional collision rate constant similar to those of other DNA-repair systems.29−34 Data from other protein− DNA-repair systems are interpreted to show that the protein binds nonspecifically to the DNA and, then, uses sliding, hopping, and/or jumping to find its target sequence or site.27,28 The target site can be located either by the enzyme flipping the target base out of the duplex DNA or by the enzyme recognizing the extrahelical target. In the case of the photolyase system, the presence of CPD bends and unwinds the dsDNA near the site of the lesion and weakens the 5′-TA base pair and, to a lesser extent, the 3′-TA base pair in the CPD.20,21,26,44,45,47 Despite obvious structural and, presumably, electrostatic changes near the CPD, photolyase apparently fails to locate the intrahelical CPD through facilitated diffusion after initial three-dimensional diffusion-controlled nonspecific binding to DNA. The recent ITC study has confirmed that nonspecific binding by photolyase to single-stranded DNA is weak (KA ≤ 200 M−1) and that electrostatic contributions of photolyase binding to the CPD are between 5 and 12% of the overall binding energy.24 This is in line with a reduced role, if any, for facilitated diffusion as pointed out by Sancar.36 The comparison with other systems suggests that photolyase may not recognize an intrahelical CPD but only an extrahelical CPD. Such a recognition mechanism has been proposed on the basis of computational studies and ITC experiments.24,25 If photolyase recognizes only the extrahelical CPD during a threedimensional diffusion-controlled search, the rate constant for this process is given by

k CPD = K flipkdiffusion

(1)

where kCPD is the binding rate constant of photolyase, Kflip is the equilibrium constant for spontaneous flipping of the CPD into an extrahelical position, and kdiffusion is the threedimensional diffusional collision rate constant. For proteins of a size similar to that of photolyase and interacting with short dsDNA, kdiffusion ≈ kon = 1.7 × 107 to 9.0 × 108 M−1 s−1.29−34 Therefore, using a kCPD of 2.5 × 105 M−1 s−1, we predict that Kflip values of 1.5 × 10−2 to 2.8 × 10−4 would be sufficient to support a purely three-dimensional search. A computational study estimated Kflip to be on the order of 5 × 10−5, and this result was interpreted in favor of photolyase recognizing an extrahelical csCPD.25 Experimentally, two NMR studies used imino proton exchange reactions to study base pair opening in csCPD-containing dsDNA. In one study, αKopen, the apparent equilibrium constant for base pair opening, was found to increase from 6.5 × 10−6 for the 5′-T·A in undamaged dsDNA to 8.9 × 10−6 for 5′-T·A of the CPD, and αKopen = 3.1 × 10−6 for 3′-T·A in undamaged dsDNA.53 The parameter α depends on the base catalyst that is used for the imino proton exchange measurements, and α ≤ 1,54 where Kopen ≥ αKopen. In a second study, it was found that Kopen increased from 8 × 10−7 for the 5′-TA in undamaged dsDNA to 2 × 10−4 for 5′-TA of the CPD, and it increased from 2.5 × 10−6 for the 3′-TA in undamaged dsDNA to 3 × 10−5 for 3′-TA of the CPD.26 In that study, α was assumed to be equal to unity for ammonia as the base catalyst. It was also pointed out that the pKa values of the CPD imino protons are not known and that an increase in their pKa by 1 pH unit due to loss of aromaticity upon CPD formation would increase Kopen by a factor of 10 to 2 × 10−3

Figure 3. Fluorescence stopped-flow kinetic traces of DNA photolyase binding to (A) csCPD-DNA1, (B) csCPD-DNA2, and (C) csCPDDNA3. The final concentrations in the mixing cell are 1.0 μM csCPDDNA and (a) 10, (b) 15, (c) 20, (d) 30, and (e) 50 μM photolyase. For the clarity of the fit () to the data (●), only every other 25th data point is shown. The inset of each graph shows the photolyase concentration dependence of the rate (■), and the binding rate constant is the slope of the linear fit () to the data.

interactions contributed to only 10% of the overall binding energy.36 Thus, they concluded that photolyase finds its target through a three-dimensional diffusion-controlled search and argued against contributions from one-dimensional or facilitated diffusion as used by the DNA-repair proteins mentioned above. The decrease of 100−4000 in the binding rate constant indicates that photolyase must use a significantly different 6182

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Biochemistry

■

and 3 × 10−4 for the 5′-TA and 3′-TA base pairs of the CPD, respectively. The authors of the NMR study argue against spontaneous flipping of the CPD into an extrahelical position and suggest that Kopen reflects flipping of the adenosine bases opposite of the CPD that gives rise to CPD imino proton exchange.26 However, the Kopen of the 5′-TA falls within our predicted range for Kflip, and, if the pKa of the imino protons would be higher in the CPD than in thymidine monomers, Kopen values for both 5′-TA and 3′-TA would fall within the range that would support recognition of an extrahelical CPD by photolyase, assuming that Kopen reflects Kflip. The preponderance of the experimental data supports the hypothesis that photolyase recognizes an extrahelical CPD. The notion that photolyase is significantly slower than other DNA-repair systems is not at odds with its efficient repair of csCPD-DNA. After photolyase has found and bound to a csCPD, it has to absorb a photon of the correct energy to catalyze the monomerization of the dimerized thymidines, which seems to be the rate-limiting step. In earlier work, we concluded that this is not a problem because substrate binding stabilizes the active form of the enzyme under physiological conditions.55 Therefore, the enzyme is not optimized for finding of the csCPD but for maintaining the enzyme−substrate complex as well as the active form of the enzyme until absorption of a blue-light photon has occurred to catalyze the repair.

Article

AUTHOR INFORMATION

Corresponding Author

*Department of Chemistry and Biochemistry, Montclair State University, 1 Normal Ave., Montclair, NJ 07043. E-mail: [email protected]. Telephone: (973) 655 3301. Present Address §

X.Z.: Moderna Therapeutics, 320 Bent St., Cambridge, MA 02141. Funding

The research was sponsored by National Science Foundation Grants MCB 0415611 (J.P.M.S.) and 0742122 (J.P.M.S.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge Drs. Richard Magliozzo and Xiangbo Zhao at Brooklyn College for their help with the stopped-flow experiments and for providing access to the stopped-flow apparatus. Part of the research was conducted at New York University (X.Z. and J.P.M.S.) and at Lafayette College (Y.M.G.).

■

REFERENCES

(1) Sancar, A. (2003) Structure and Function of DNA Photolyase and Cryptochrome Blue-Light Photoreceptors. Chem. Rev. 103, 2203− 2237. (2) Brettel, K., and Byrdin, M. (2010) Reaction Mechanism of DNA Photolyase. Curr. Opin. Struct. Biol. 20, 693−701. (3) Kneuttinger, A. C., Kashiwazaki, G., Prill, S., Heil, K., Müller, M., and Carell, T. (2014) Formation and Direct Repair of UV-induced Dimeric DNA Pyrimidine Lesions. Photochem. Photobiol. 90, 1−14. (4) Taylor, J.-S. (1994) Unraveling the Molecular Pathway from Sunlight to Skin Cancer. Acc. Chem. Res. 27, 76−82. (5) Mouret, S., Baudouin, C., Charveron, M., Favier, A., Cadet, J., and Douki, T. (2006) Cyclobutane pyrimidine dimers are predominant DNA lesions in whole human skin exposed to UVA radiation. Proc. Natl. Acad. Sci. U. S. A. 103, 13765−13770. (6) MacFarlane, A. W., IV, and Stanley, R. J. (2003) Cis-Syn Thymidine Dimer Repair by DNA Photolyase in Real Time. Biochemistry 42, 8558−8568. (7) Kao, Y.-T., Saxena, C., Wang, L., Sancar, A., and Zhong, D. (2005) Direct observation of thymine dimer repair in DNA photolyase. Proc. Natl. Acad. Sci. U. S. A. 102, 16128−16132. (8) Thiagarajan, V., Byrdin, M., Eker, A. P. M., Muller, P., and Brettel, K. (2011) Kinetics of cyclobutane thymine dimer splitting by DNA photolyase directly monitored in the UV. Proc. Natl. Acad. Sci. U. S. A. 108, 9402−9407. (9) Park, H.-W., Kim, S.-T., Sancar, A., and Deisenhofer, J. (1995) Crystal structure of DNA photolyase from Escherichia coli. Science 268, 1866−1872. (10) Vande Berg, B. J., and Sancar, G. B. (1998) Evidence for Dinucleotide Flipping by DNA Photolyase. J. Biol. Chem. 273, 20276− 20282. (11) Sanders, D. B., and Wiest, O. (1999) A Model for the Enzyme Substrate Complex of DNA Photolyase and Photodamaged DNA. J. Am. Chem. Soc. 121, 5127−5134. (12) Antony, J., Medvedev, M., and Stuchebrukhov, A. A. (2000) Theoretical Study of Electron Transfer between the Photolyase Catalytic Cofactor FADH- and DNA Thymine Dimer. J. Am. Chem. Soc. 122, 1057−1065. (13) MacFarlane, A. W., IV, and Stanley, R. J. (2001) Evidence of Powerful Substrate Electric Fields in DNA Photolyase: Implications for Thymidine Dimer Repair. Biochemistry 40, 15203−15214. (14) Christine, K. S., MacFarlane, A. W., Yang, K., and Stanley, R. J. (2002) Cyclobutanepyrimidine Dimer Base Flipping by DNA Photolyase. J. Biol. Chem. 277, 38339−38344.

■

CONCLUSION Our results show that dsDNA can be labeled with a fluorophore to study the binding of photolyase to csCPD-containing DNA. However, one has to pay careful attention to the location of the fluorophore with respect to the csCPD lesion to avoid steric interference from the fluorophore during the binding process. Neither the location of the csCPD on the undecamer dsDNA nor the presence of the fluorophore has any significant effect on the thermal stability of the DNA as determined by its melting temperature. When we omit the data that appear to be affected by steric interference from the fluorophore, we find that the association constant of photolyase is independent of the location of the csCPD whether in the center or at the terminus of the undecamer DNA. The Ka value of ∼2 × 106 M−1 is in good agreement with the literature for similar substrates. In contrast, the kinetics of binding of photolyase to the csCPD undecamer DNA is sensitive to the relative position of the csCPD with respect to the fluorophore. For the binding rate constant, we find a kon of ∼2.5 × 105 M−1 s−1, which is slightly slower than rate constants that were obtained through indirect measurements in vivo in E. coli cells and in vitro with plasmid DNA.35,36 Most importantly, the photolyase binding rate constant is 100− 4000 times slower than that for other DNA-repair systems. These other proteins bind nonspecifically to DNA with a rate constant that is limited by three-dimensional diffusion coefficients and rapidly locate their target through facilitated diffusion in the form of sliding, hopping, and/or jumping along the DNA without completely dissociating. It appears that photolyase does not follow such a search process and instead may rely on a three-dimensional search for an extrahelical csCPD, which spontaneously flips in and out of the double helix. Experiments to further test this proposed mechanism are underway. 6183

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry (15) Torizawa, T., Ueda, T., Kuramitsu, S., Hitomi, K., Todo, T., Iwai, S., Morikawa, K., and Shimada, I. (2004) Investigation of the Cyclobutane Pyrimidine Dimer (CPD) Photolyase DNA Recognition Mechanism by NMR Analysis. J. Biol. Chem. 279, 32950−32956. (16) Schelvis, J. P. M., Ramsey, M., Sokolova, O., Tavares, C., Cecala, C., Connell, K., Wagner, S., and Gindt, Y. M. (2003) Resonance Raman and UV-Vis Spectroscopic Characterization of FADH• in the Complex of Photolyase with UV-Damaged DNA. J. Phys. Chem. B 107, 12352−12362. (17) Kapetanaki, S. M., Ramsey, M., Gindt, Y. M., and Schelvis, J. P. M. (2004) Substrate Electric Dipole Moment Exerts a pH-Dependent Effect on Electron Transfer in Escherichia coli Photolyase. J. Am. Chem. Soc. 126, 6214−6215. (18) Murphy, A. K., Tammaro, M., Cortazar, F., Gindt, Y. M., and Schelvis, J. P. M. (2008) Effect of Cyclobutane Cytidine Dimer on the Properties of Escherichia coli DNA Photolyase. J. Phys. Chem. B 112, 15217−15226. (19) Mees, A., Klar, T., Gnau, P., Hennecke, U., Eker, A. P. M., Carell, T., and Essen, L.-O. (2004) Crystal Structure of a Photolyase Bound to a CPD-Like DNA Lesion After in situ Repair. Science 306, 1789−1793. (20) McAteer, K., Jing, Y., Kao, J., Taylor, J.-S., and Kennedy, M. A. (1998) Solution-state Structure of a DNA Dodecamer Duplex Containing a Cis-Syn Thymine Cyclobutane Dimer, the Major UV Photoproduct of DNA. J. Mol. Biol. 282, 1013−1032. (21) Park, H., Zhang, K., Ren, Y., Nadji, S., Sinha, N., Taylor, J.-S., and Kang, C. (2002) Crystal structure of a DNA decamer containing a cis-syn thymine dimer. Proc. Natl. Acad. Sci. U. S. A. 99, 15965−15970. (22) Roberts, R. J., and Cheng, D. (1998) Base Flipping. Annu. Rev. Biochem. 67, 181−198. (23) Stivers, J. T., and Jiang, Y. L. (2003) A Mechanistic Perspective on the Chemistry of DNA Repair Glycosylases. Chem. Rev. 103, 2729− 2759. (24) Wilson, T. J., Crystal, M. A., Rohrbaugh, M. C., Sokolowsky, K. P., and Gindt, Y. M. (2011) Evidence from Thermodynamics that DNA Photolyase Recognizes a Solvent-Exposed CPD Lesion. J. Phys. Chem. B 115, 13746−13754. (25) O’Neil, L. L., Grossfield, A., and Wiest, O. (2007) Base Flipping of the Thymine Dimer in Duplex DNA. J. Phys. Chem. B 111, 11843− 11849. (26) Wenke, B. B., Huiting, L. N., Frankel, E. B., Lane, B. F., and Núñez (2013) Base Pair Opening in a Deoxynucleotide Duplex Containing a cis-syn Thymine Cyclobutane Dimer Lesion. Biochemistry 52, 9275−9285. (27) Halford, S. E. (2009) An End to 40 Years of Mistakes in DNAProtein Association Kinetics? Biochem. Soc. Trans. 37, 343−348. (28) Friedman, J. I., and Stivers, J. T. (2010) Detection of Damaged DNA Bases by DNA Glycosylase Enzymes. Biochemistry 49, 4957− 4967. (29) Barkley, M. D. (1981) Salt Dependence of the Kinetics of the lac Repressor-Operator Interaction: Role of Nonoperator Deoxyribonucleic Acid in the Association Reaction. Biochemistry 20, 3833−3842. (30) Winter, R. B., Berg, O. G., and von Hippel, P. H. (1981) Diffusion-Driven Mechanisms of Protein Translocation on Nucleic Acids 3. The Escherichia coli lac Repressor-Operator Interaction: Kinetic Measurements and Conclusions. Biochemistry 20, 6961−6977. (31) Hopkins, B. B., and Reich, N. O. (2004) Simultaneous DNA Binding, Bending, and Base Flipping. J. Biol. Chem. 279, 37049−37060. (32) Krosky, D. J., Song, F., and Stivers, J. T. (2005) The Origins of High-Affinity Enzyme Binding to an Extrahelical DNA Base. Biochemistry 44, 5949−5959. (33) Porecha, R. H., and Stivers, J. T. (2008) Uracil DNA Glycosylase Uses DNA Hopping and Short-Range Sliding to Trap Extrahelical Uracils. Proc. Natl. Acad. Sci. U. S. A. 105, 10791−10796. (34) Iakoucheva, L. M., Walker, R. K., van Houten, B., and Ackerman, E. J. (2002) Equilibrium and Stopped-Flow Kinetic Studies of Fluorescently Labeled DNA Substrates with DNA Repair Proteins XPA and Replication Protein A. Biochemistry 41, 131−143.

(35) Harm, W. (1970) Analysis of Photoenzymatic Repair of UV Lesions in DNA by Single Light Flashes. V. Determination of the Reaction Rate Constants in E. coli Cells. Mutat. Res., Fundam. Mol. Mech. Mutagen. 10, 277−290. (36) Sancar, G. B., Smith, F. W., Reid, R., Payne, G., Levy, M., and Sancar, A. (1987) Action Mechanism of Escherichia coli DNA Photolyase. J. Biol. Chem. 262, 478−485. (37) Gindt, Y. M., Vollenbroek, E., Westphal, K., Sackett, H., Sancar, A., and Babcock, G. T. (1999) Origin of the Transient Electronic Paramagnetic Resonance Signals in DNA Photolyase. Biochemistry 38, 3857−3866. (38) Förster, T. (1948) Zwischenmolekulare Energiewanderung und Fluoreszenz. Ann. Phys. 437, 55−75. (39) Stryer, L., and Haugland, R. P. (1967) Energy Transfer. Spectroscopic Ruler. Proc. Natl. Acad. Sci. U. S. A. 58, 719−726. (40) Wang, B., and Jorns, M. S. (1989) Reconstitution of Escherichia coli DNA Photolyase with Various Folates. Biochemistry 28, 1148− 1152. (41) Banerjee, S. K., Christensen, R. B., Lawrence, C. W., and LeClerc, J. E. (1988) Frequency and Spectrum of Mutations Produced by a Single cis-syn Thymine-Thymine Cyclobutane Dimer in a SingleStranded Vector. Proc. Natl. Acad. Sci. U. S. A. 85, 8141−8145. (42) Cantor, C. R., Warshaw, M. M., and Shapiro, H. (1970) Oligonucleotide Interactions III: Circular Dichroism Studies of the Formation of Deoxyoligonucleotides. Biopolymers 9, 1059−1077. (43) Taylor, J.-S., Garrett, D. S., Brockie, I. R., Svoboda, D. L., and Telser, J. (1990) 1H NMR Assignment and Melting Temperature Study of Cis-Syn and Trans-Syn Thymine Dimer Containing Duplexes of d(CGTATTATGC)•d(GCATAATACG). Biochemistry 29, 8858− 8866. (44) Yang, K., and Stanley, R. J. (2006) Differential Distortion of Substrate Occurs When It Binds to DNA Photolyase: A 2Aminopurine Study. Biochemistry 45, 11239−11245. (45) Yang, K., Matsika, S., and Stanley, R. J. (2007) 6MAP, a Fluorescent Adenine Analogue, Is a Probe of Base Flipping by DNA Photolyase. J. Phys. Chem. B 111, 10615−10625. (46) Nazarenko, I., Pires, R., Lowe, B., Obaidy, M., and Rashtchian, A. (2002) Effect of Primary and Secondary Structure of Oligodeoxyribonucleotides on the Fluorescenct Properties of Conjugated Dyes. Nucleic Acids Res. 30, 2089−2195. (47) Yang, K., and Stanley, R. S. (2008) The Extent of DNA Deformation in DNA Photolyase-Substrate Complexes: A Solution State Fluorescence Study. Photochem. Photobiol. 84, 741−749. (48) Husain, I., Sancar, G. B., Holbrook, S. R., and Sancar, A. (1987) Mechanism of Damage Recognition by Escherichia coli DNA Photolyase. J. Biol. Chem. 262, 13188−13197. (49) Husian, I., and Sancar, A. (1987) Binding of E. coli photolyase to a defined substrate containing a single T< > T dimer. Nucleic Acids Res. 15, 1109−1120. (50) Kim, S.-T., and Sancar, A. (1991) Effect of base, pentose, and backbone structures on binding and repair of pyrimidine dimers by Escherichia coli DNA photolyase. Biochemistry 30, 8623−30. (51) Jordan, S. P., and Jorns, M. S. (1988) Evidence for a singlet intermediate in catalysis by Escherichia coli DNA photolyase and evaluation of substrate binding determinants. Biochemistry 27, 8915− 8923. (52) Jordan, S. P., Alderfer, J. L., Chanderkar, L. P., and Jorns, M. S. (1989) Reaction of Escherichia coli and yeast photolyases with homogenous short-chain oligonucleotide substrates. Biochemistry 28, 8149−8153. (53) Bang, J., Kang, Y.-M., Park, C.-J., Lee, J.-H., and Choi, B.-S. (2009) Thermodynamics and Kinetics for Base Pair Opening in the DNA Decamer Duplexes Containing Cyclobutane Pyrimidine Dimer. FEBS Lett. 583, 2037−2041. (54) Kochoyan, M., Leroy, J. L., and Guéron, M. (1987) Proton Exchange and Base-pair Lifetimes in a Deoxy-duplex Containing a Purine-Pyrimidine Step and in the Duplex of Inverse Sequence. J. Mol. Biol. 196, 599−609. 6184

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185

Article

Biochemistry (55) Gindt, Y. M., Schelvis, J. P. M., Thoren, K. L., and Huang, T. H. (2005) Substrate Binding Modulates the Reduction Potential of DNA Photolyase. J. Am. Chem. Soc. 127, 10472−10473.

6185

DOI: 10.1021/acs.biochem.5b00927 Biochemistry 2015, 54, 6176−6185